Monday, August 3rd, 2015

Time	Event
12:00p	Pollution in China Makes Free Cooling Difficult for Baidu China’s notoriously high air pollution levels are a well-documented public-health issue. But pollution also has other less talked about effects. One of them is on the efficiency of data centers in the country. Pulling outside air into a data center to cool equipment to reduce energy used by mechanical cooling systems has been one of the most effective ways to increase the facility’s energy efficiency. The high concentration of pollutants in the atmosphere, however, makes “free cooling” in China data centers not as easily obtainable as it is in the US, for example, where the use of outside air to supplement data center cooling capacity is commonplace. Air pollution in China has resulted in higher IT equipment failure rates for Baidu, the internet giant that’s the country’s answer to Google. The problem of pollution has driven a lot of research and development focus on data center cooling technologies at the company. Those research efforts take place both in China and in Baidu’s Silicon Valley offices, Ali Heydari, the company’s chief architect, said. Heydari talked about the challenges of operating data centers in China at the DCD Internet conference in San Francisco Friday. The highest concentrations of air pollutants are found in the eastern portion of China. “Most of our data centers are actually located in this area,” Heydari said. “That’s a major challenge for us.” Baidu has three data centers in Beijing, whose smoggy images have become emblematic of China’s pollution problem, and one in Shanxi, a province just west of the capital. Beijing Smog The pollutants include sulphur dioxide, nitrogen oxide, and PM 2.5, which is shorthand for fine particulate matter. PM 2.5, generated by vehicle exhaust, coal power plants, or fires, is what creates the visible haze in polluted cities and causes many of the human health problems associated with air pollution. Concentration of PM 2.5 is a common air quality metric. The US Environmental Protection Agency uses an Air Quality Index system, where quality ranges from 0 (cleanest) to 500. The AQI for PM 2.5 in Beijing on early Saturday morning local time was over 150, which the EPA considers “unhealthy.” Around the same time, the AQI for PM 2.5 in Ashburn, Virginia, which has one of the biggest concentrations of data centers in the US, was about 30, considered “good.” The Chinese government recently claimed major improvements in air quality in Beijing, reportedly due to environmental policies of the last year, but problems persist, causing issues for data center operators. Pollution causes salts to accumulate in data center air conditioning systems, and exposure to gas pollutants significantly increases electronics corrosion rates, Heydari said. Pollution Drives Cooling Innovation One of the technological solutions his team has been looking into is air scrubbing. They have been testing a variety of scrubbing methods, such as water or chemical spraying or filtration. The idea is to integrate a scrubbing mechanism into the free cooling system. The liquid spraying approach is effective but requires a lot of dehumidification. They have also created cooling designs that don’t rely on outside air but increase efficiency otherwise. One such design that is showing a lot of promise is a “bottom cooling unit.” Essentially, instead of a centralized chiller plant, the heat-exchanger coils are placed directly underneath the IT racks. The racks themselves have enclosures around them, and cold air is pushed into the enclosures, ensuring that all cold air that’s created makes it to the hardware. “It’s basically a self-contained data center,” Heydari said about the enclosure. “It has its own cold aisle; it has its own hot aisle.” The bottom cooling units don’t require a raised floor, however, they are most efficient when installed in a raised-floor environment. The system automatically adjusts cooling capacity based on the hardware density installed in the rack. Because cold-air supply is localized, it is more efficient and eliminates the risk of widespread cooling failure. It is a liquid-cooling solution, but because the actual heat exchanger is under the rack, there’s no danger of damage to IT equipment by a water leak. Innovating for Web Scale Like its web-scale peers in the US, Baidu spends a lot of resources on R&D to increase the efficiency of its hardware and data centers. The company has been involved in Facebook’s Open Compute Project and has designed its own servers and data center racks that use some OCP design elements. Heydari, who in the past has worked as a senior hardware engineer at Facebook and Twitter, says Baidu’s Scorpio server designs are more energy efficient and yield higher power densities than OCP designs. The designs aren’t Baidu’s alone. Project Scorpio was started by a group of China’s largest data center operators, which also included Tencent, Alibaba, and China Telecom. Baidu is also an early adopter of servers powered by ARM processors, the low-power alternative to Intel’s x86 architecture, FPGAs (field-programmable gate arrays), and GPU accelerators, according to Heydari. A Long-Term Issue for Growing Data Center Market Like other web-scale data center operators, Baidu has had to innovate out of necessity since typical hardware vendors have not traditionally designed equipment for infrastructure of such scale. In China, air quality happened to create the necessity to think differently about data center cooling. This and other challenges in that market – some of the other major challenges there are high energy costs and relatively low rack power densities in colocation facilities – are important to address, since much of the growth in the data center sector in the coming years is going to occur in China, one of the world’s fastest-growing cloud services markets. (Comment on this)
3:00p	Load Shedding: Five Reasons it Matters for Your Applications Kris Beevers is the CEO & Founder of NSONE. The concept of load shedding in web infrastructure is a pretty simple one. If a particular system in an application’s delivery stack is overloaded, “shed” some of the load so the system can at least continue to provide service for a subset of requests, rather than falling over and dying. In today’s application stacks, which are often multi-faceted and distributed across data center environments for resiliency and performance, load shedding plays a powerful role in efficient application delivery, preventing outages and maximizing the efficiency of the infrastructure. To support modern application architectures, today’s most advanced DNS and traffic management providers build load-shedding capabilities directly into their services, making it possible for any application to easily take advantage of this powerful technology. Here are five ways that load shedding can ensure your distributed application is delivered reliably and efficiently to your users. Load Shedding Prevents Cascading Failures A cascading failure is an incident that starts in one system or data center in your architecture and causes a chain reaction of failures in your other systems and data centers. For example, if a load balancer fails in one of your delivery facilities, a naive failover approach might be to shift all that traffic to the “next best” data center, say, the next closest. The sudden flood of failover traffic may overwhelm the load balancer in the failover data center, and so on. A traffic management platform that supports load shedding can take in data from your systems, like system load metrics or connection counts from your load balancers, and ensure none of your systems are pushed beyond their limits. With load shedding, when a load balancer in one of your data centers fails, the bulk of its traffic can be shifted over to the next closest data center, up to a load watermark or threshold for that secondary facility. After that, the rest of the traffic can be shifted to a tertiary data center to avoid overloading the secondary one. Load shedding can cascade your traffic across a number of facilities and avoid overloading any of them. Load Shedding Prevents ‘Flapping’ The “brute force” way to deal with an overloaded data center is to shut it off and handle the traffic elsewhere. This binary approach is ill-suited for today’s traffic spikes, and in fact causes painful “flapping” in some situations. For example, if DC1 gets more traffic than it can handle and is brought offline, traffic is routed elsewhere. When DC1 recovers and comes back online, traffic is shifted back, quickly overloading DC1 again. Rinse and repeat. Instead of a binary “on-off” approach to overload situations, load shedding enables you to use a data center to capacity but not beyond, bleeding away the traffic that causes an overload to your other facilities. There is no need to shut off traffic altogether, as overflow traffic that causes delays or failure is easily shifted elsewhere, keeping all systems humming along at capacity. Load Shedding Makes Efficient Use of Infrastructure In a distributed application, there’s a push and pull between optimizing for performance and optimizing for heavy traffic. If you distribute your spend (and traffic) effectively across a number of data centers and push your content to the “edge,” users will then see reduced response times and have a better experience with your application. Spreading your infrastructure too thin can be disastrous when there is a localized traffic spike or attack, causing flapping or cascading failures. With load shedding, it’s okay to run lean in your edge facilities and build mostly for the common case, optimizing delivery performance. Abnormally high and localized traffic is smoothly transitioned to your other facilities in case of a spike. Load Shedding = Infinite Capacity, or at Least, Graceful Failures Even if you’re stuck in a single data center, or with a limited infrastructure budget, load shedding can help you deal with large traffic spikes without falling over entirely. When your global capacity is maxed out, load shedding can shift overflow traffic to a lightweight, “failwhale” page hosted by a third party or in on-demand infrastructure. You’ll continue to provide good service to most of your users, and a better failure experience to the rest. Load Shedding Plays Nicely with Other Traffic Management Tools Modern DNS services provide many different advanced traffic management technologies in addition to load shedding, such as geographic routing, network-based fencing, latency-based routing and more. Chances are, you’ll want to take advantage of these tools to distribute your “nominal” traffic across your data centers and leverage load shedding to help deal with traffic spikes or other extraordinary situations. The good news is, platforms with advanced traffic management tools, like load shedding, make it easy to mix and match routing algorithms to achieve complex behaviors, helping you improve your application’s performance, maximize your infrastructure ROI and delight your users. Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library. (Comment on this)
4:12p	Financial Firms’ Cloud Security Practices Maturing: Report This article originally appeared at The WHIR Financial services firms are aggressively using encryption and tokenization to maintain data security while adopting cloud computing, according to a new report fromCipherCloud. The report shows that financial companies employ a rapidly maturing approach to cloud, with almost every one storing personally identifiable information (PII) in the cloud. CipherCloud surveyed employees of over 50 global banking and financial services firms from North America, Europe, Asia Pacific and Latin America, asking them about strategies used to protect personally identifiable information in the cloud. The results inform the company’s Q2 2015 Global Cloud Data Security Report. While almost all companies surveyed store some PII in the cloud, which data and how it is stored varies. Highly sensitive data is only held in the cloud by 33 percent, while the cloud is used by 47 percent to process personal financial data and 53 percent to store confidential business data. “As financial services adopt the cloud, strict compliance regulations and corporate policies push them to be early adopters of security technologies,” said Pravin Kothari, founder and CEO, CipherCloud. “At the same time, the influence of cloud has upped the ante for financial services firm CISOs and their teams. As these companies increase their cloud adoption, they are building data protection in the cloud with the help of innovative encryption and tokenization technologies. Both regulatory scrutiny and the pace of data breaches compel the increased protection of their sensitive information.” Tokenization is used by 40 percent of firms for protecting highly sensitive PII, and the use of both tokens and strong encryption increases with the sensitivity of the data being protected. Encryption is favored over tokenization for less sensitive data, likely to ease frequent search and sort functions, the report says. Sixty-four percent use searchable encryption, balancing searchability and security strength to support workflows where data references or indexes other information. The report also stresses the importance of format preserving encryption for data like email addresses, which is the method used by 91 percent of respondents. More than a billion records with PII were leaked in 2014, according to an IBM report. A March report sponsored by CipherCloud showed that financial institutions were working towards cloud strategies, but less than 30 percent in the US had one in place. Evidence that financial institutions would spend to address cybersecurity in the wake of a series of bank hacks began to role in with a Temenos survey last September, and CipherCloud will offer its data protection and compliance solutions in a bid for a share of the growing market. This first ran at http://www.thewhir.com/web-hosting-news/financial-firms-cloud-security-practices-maturing-report (Comment on this)
4:28p	Vendors: OpenStack Buying Decisions Moving From IT to Line of Business OpenStack recently had its fifth anniversary. The open source project now has thousands of contributors, and it’s time to stop talking about the building blocks and start talking about the problems cloud solves. “I can’t wait until OpenStack takes the back burner in terms of conversation,” said Cisco Cloud director Scott Sanchez. “It’s about ‘what business problem am I solving?’ That’s where I have always wanted to see this go.” Sanchez came to Cisco from the company’s Metacloud acquisition. Metacloud provides hosted private cloud – one of many legs in Cisco’s Intercloud strategy. Sanchez was part of the original OpenStack team at Rackspace. “During the first two years at Rackspace, my job description was ‘help OpenStack win’,” said Sanchez. Sanchez said that who’s buying cloud has changed. While large companies are budgeting for OpenStack, that budget is not coming from IT. It’s coming from line of business. Industry insiders love to talk about the building blocks of OpenStack, and for good reasons. An increasing number of companies have a stake in its success. However, where this goes wrong is making project details the primary message to customers. The conversation should instead be how business problems are being solved. Sanchez gave an example of a large bank customer. “We deal with two sides of the company: we work with the traditional data center team and we’re also working with their global head of greenfield applications,” he said. “That person sits inside the line of business, outside of IT entirely, reporting to the actual people making money for the bank. They’re the ones funding the necessity for private cloud targets.” Internap Says Conversation Has Changed Data center service provider Internap has evolved over the years, but VP and general manager of cloud Satish Henachandran said the company has stayed true to the original vision: focusing on technology problems where performance matters. Internap was an early adopter of OpenStack. Henachandran said its first OpenStack offering brought to market, Swift object storage, “didn’t do as well as we intended, but it was certainly a proving ground.” On its fifth anniversary, Henachandran believes, OpenStack is now a fundamental building block that has taken a backseat in conversations. “We shouldn’t be everything for everyone. The solution set in terms of hybrid is around ‘What specific customer problem are we trying to solve?’” said Henachandran. “The answer is either growing revenue or reducing cost as it maps to line of business-oriented or IT-oriented workload.” Internap has since added an OpenStack public cloud service which isn’t positioned against other public clouds but rather as a piece of a larger hybrid package. “It’s an add-on versus a primary offering,” said Henachandran. iWeb, a Montreal-based hosting provider Internap acquired in 2013, is also independently looking at OpenStack as a strategic initiative. On OpenStack’s fifth birthday, calling OpenStack “boring” is a big compliment. OpenStack is so ingrained in the foundation of many cloud initiatives that it now feels less like a gamble for many customers. OpenStack is ubiquitous and the conversation is changing. (Comment on this)
5:20p	Analytics in DCIM Software Drives Data Center Efficiencies Given the rising complexity of data center environments, keeping track of what piece of IT infrastructure is being used for what purpose has moved beyond the ability of most IT. In fact, the days when IT organizations can get by using spreadsheets to keep accurate inventory of their IT infrastructure are all but over. At the Data Center World conference in National Harbor, Maryland this September, Brad Beamish, senior technical consultant for Cormant, a provider of data center infrastructure management software, will explain why DCIM software is not only a critical tool for gaining visibility into the data center environment, but also, just as importantly, for generating analytics insights in a way IT operations teams can comprehend. “Data speaks its own language, but analytics speaks your language,” said Beamish. “Matching the right analytics with your business goals is where you uncover the opportunities for data center transformation.” Key DCIM software decisions that have to be upfront, he said, include figuring out what is actually worth monitoring inside your environment in the first place. Once that’s determined, the next challenge is to turn that data into actionable intelligence. Armed with that information data center managers can make the right decisions concerning where any given application workloads should be running at any given time. That not only serves to increase utilization rates of IT infrastructure inside the data center; it ultimately serves to reduce energy costs and management overhead across the entire lifespan of the data center facility. For more information, sign up for Data Center World National Harbor, which will convene in National Harbor, Maryland, on September 20-23, 2015, and attend Brad’s session titled “Data, Data and More Data: Key Data Center Analytics to Track in your DCIM Tool Now.” (Comment on this)
6:24p	LuxConnect Data Center in Luxembourg Nears Launch Luxembourg data center provider LuxConnect is putting the final touches on its fourth data center in the country. The data center is designed for high resiliency and will use clean energy sourced in Norway. The provider has a Tier IV design certification for the facility from the Uptime Institute, and said it plans to also go through the process for certifying the constructed facility. It is easier to buy renewable electricity at the large scale that data centers need in northern Europe than it is in other parts of the world, including the US. Renewable energy is abundant in the region, and relatively accessible because of deregulated energy markets there, and because electrical grids across multiple northern European countries are interconnected. For customers in the US, buying energy even across state lines is often very complicated because of the way energy regulation is structured. Tier IV is the highest data center infrastructure reliability level recognized by Uptime. In addition to the new facility in Bettembourg, LuxConnect has Tier IV design certification for two older data centers. The newest data center, as well as one other older one in Bissen, also have Tier II design certification. None of LuxConnect’s existing data centers have a Tier certification for constructed facility. This is an important distinction, since many data center providers get design certification but not the constructed facility certification, which means they haven’t had Uptime verify that the data center was actually built to the design that was certified. The design v. constructed facility certification issue is a thorny one, some data center operators claiming that the practice of getting the former without following through with the latter amounts to abuse of the rating system and serves to mislead customers about a data center’s actual reliability level. The controversy led Uptime to revise the system recently. While design certifications are still available to all data center operators elsewhere around the world, US data center service providers may no longer get them. Last year, Uptime also said operators that received design certifications will lose them if they don’t get constructed facility certifications within two years. LuxConnect is planning to follow-up with a facility certification for at least its newest facility. “Uptime Institute will do the review on the tests we have done together, and we are hoping to be granted full certification by the time the site opens later this summer,” LuxConnect CEO Roger Lampach said in a statement. The data center will have two redundant 10 MVA electrical feeds. Half of it has been fully built out (the half that’s designed to Tier IV standards), and half consists of shell space for custom designs for potential clients. It will have about 60,000 square feet of data center space, according to the LuxConnect website. (Comment on this)
7:20p	Mesos Update Improves Data Center Resource Management Popular open source software for large cluster management Apache Mesos has been updated with new functionality. The 0.23.0 release has new features around managing resources at large scale, expanding its capabilities in an automated data center. Mesos plays an important role in modern data centers. Fewer people manage an increasing number of servers companies that provide services at web scale services, acting as conductors of a distributed systems symphony. Mesos abstracts disparate resources, bundling them into virtual pools. The new features are around better predictions and automation of resource utilization, including oversubscription, persistent volumes, and dynamic reservations. Mesos is used to power the backend of some heavy-duty services, such as Apple’s Siri and Twitter. Mesos is credited for killing the “Fail Whale” that was often seen in Twitter’s early years. Mesos is also the kernel of Mesosphere’s Data Center Operating System (DCOS), which gains the new functionality as well. The new oversubscription feature allows taking advantage of temporarily unused resources to execute best-effort tasks on lower-priority jobs, such as background analytics and video or image processing. “High-priority user-facing services are typically provisioned on large clusters for peak load and unexpected load spikes. This means that for most of the time the cluster is under-utilized,” wrote Mesosphere data center application architect Michael Hausenblas. “In the context of the DCOS, this means that you will be able to get even more bang for your bucks: more applications with the same hardware investments or lower TCO.” Oversubscription adds two new slave components: resource estimator and a quality of service controller. The resource estimator hooks into resource monitor to get usage statistics and identify the amount of oversubscribed resources. The QoS controller informs when corrective actions need to be made. Mesos also now provides a mechanism to create a persistent volume for stateful services. Previously, distributed file systems, message queues, and other functions had to use out-of-band techniques. “With persistent volumes, these stateful services have now become first-class citizens, making the life easier for DCOS service developers and providing a more robust experience for DCOS end-users,” wrote Hausenblas. The 0.23.0 release also includes experimental support for dynamic reservations. Dynamic reservations means a human operator doesn’t need to specify fixed, static reservations on startup – instead resources can be reserved as they are being offered without a slave restart. “This new feature means for DCOS operators less manual interventions when business requirements change priorities concerning the cluster usage,” wrote Hausenblas. Other features include SSL encryption in Mesos between master, slave, and services, and per-container network isolation. “Network isolation prevents a single container from exhausting the available network ports, consuming an unfair share of the network bandwidth or significantly delaying packet transmission for others,” wrote Hausenblas. (Comment on this)

<< Previous Day

2015/08/03 [Calendar]

Next Day >>